Skip to Main content Skip to Navigation
Theses

OPTIMISATION DE LA RECONSTRUCTION COMPLÈTE 3D EN TOMOGRAPHIE PAR ÉMISSION DE POSITONS DU PETIT ANIMAL PAR MODÉLISATION MONTE-CARLO DE LA MATRICE SYSTÈME

Abstract : The quality of PET images depends strongly on the image reconstruction method. Statistical iterative methods help to reduce the impact of the physical artefact sources on the image, since they can incorporate a model of the system response. We implemented fully 3-D reconstruction technique based on an accurate description of the system response to reconstruct listmode data. The main feature of our method, named SIMALIR, is the off-line calculation of the system probability matrix by Monte Carlo simulations. It was applied on the Philips MOSAICTM animal PET system. The tomograph was simulated using GATE toolkit. A comprehensive evaluation of the MOSAICTM model was performed in order to describe reliably the system response function. The SIMALIR method was compared with FBP, RAMLA-3D and MLEM using phantoms and mice data. The results show that SIMALIR improves considerably the images quality compared to classical PET reconstruction algorithms.
Complete list of metadatas

Cited literature [183 references]  Display  Hide  Download

https://tel.archives-ouvertes.fr/tel-00536301
Contributor : Charbel Merheb <>
Submitted on : Thursday, November 18, 2010 - 7:40:58 PM
Last modification on : Thursday, December 10, 2020 - 3:42:39 AM
Long-term archiving on: : Saturday, February 19, 2011 - 2:42:24 AM

Identifiers

  • HAL Id : tel-00536301, version 1

Citation

Charbel Merheb. OPTIMISATION DE LA RECONSTRUCTION COMPLÈTE 3D EN TOMOGRAPHIE PAR ÉMISSION DE POSITONS DU PETIT ANIMAL PAR MODÉLISATION MONTE-CARLO DE LA MATRICE SYSTÈME. Biophysique [physics.bio-ph]. Université Paris Sud - Paris XI, 2007. Français. ⟨tel-00536301⟩

Share

Metrics

Record views

711

Files downloads

3190